Playbook · updated September 1, 2026

Reducing unplanned downtime, in the order that pays

In shortPlants that cut downtime 20–50% follow the same sequence: quantify the cost per line (most underestimate it), rank assets by actual loss history, pick a sensing philosophy per asset class (vibration hardware for rotating assets, process-signal drift for the rest), and wire predictions into the existing CMMS workflow so alerts become work orders. Technology last, workflow first.

1 · Put a number on it

Run the downtime cost calculator per line, with contribution margin, not revenue. The Fortune Global 500 loses an estimated ~11% of revenues to unplanned downtime (Siemens, The True Cost of Downtime); most plants that do this exercise for the first time find a number nobody was accountable for.

2 · Rank assets by loss, not by criticality theory

Twelve months of stoppage history, sorted by margin lost. The top ten assets usually cover 60–80% of the pain. Those ten are the program scope — resist plant-wide rollouts on day one.

3 · Choose sensing per asset class

Rotating assets (pumps, fans, gearboxes, compressors): dedicated vibration sensing with diagnostic models — deepest prescriptions, per-asset hardware cost. Everything else (and rotating assets you won't instrument yet): drift detection on the process signals you already historize — no hardware, catches precursors that never vibrate. Most plants need both; the loss ranking tells you the mix.

4 · Wire predictions into the workflow

A prediction that doesn't become a work order is a demo. Integrate alerts into the CMMS with an owner, a due date and a feedback field ('was the fault real?') — that feedback loop is what keeps models honest and technicians engaged. Programs run this way report 20–50% of unplanned stops recovered; programs run as dashboards report screenshots.

Run the numbersPrice this loss with the What does unplanned downtime cost you? calculator, then take the ROI workbook into the budget meeting.